SenseCare - Health Prediction and Wellness - Use Case
A public hospital specializing in Neuroscience is working with a cluster of social service agencies to develop an emotion AI tool to identify signs of non-motor Stage 1 and Stage 2 symptoms for Depression, Anxiety, Cognitive decline, Psychosis, insomnia and Hallucinations. The hospital also wants to develop an emotion AI detection tool to identify signs indicating a pre-stroke condition as well as post-stroke recovery. The system should also be able to handle pain detection. The hospital wants the system for occupational therapy and predictions of future health.
- Create a non-intrusive tool to monitor patients for early detection of changes in mental well-being. The goal is to aid self-help management and behavioural therapy. The system will be used in rehabilitation as well as pain detection for stroke victims suffering acute pain,. The underlying purpose is to develop an early warning indicator for medical conditions such as Dementia and Alzheimer’s Disease.
- To integrate emotion recognition AI with more patient databases and sharing models, in order to develop more medical service modalities that can monitor and predict the health status of patients
- Dementia or Alzheimer’s Disease is a complex systemic disorder with many nonmotor symptoms including neuropsychiatric features such as cognitive impairment and psychosis. The objective is to aid prevention with tailored combinations of lifestyle and disease modification for long-term neurological and mental therapies.
- SenseCare’s algorithm with deep learning automated facial expression recognition can discern emotions with high accuracy and ability to identify feelings that might otherwise be concealed. With the addition of multimodal speech, physiological signals and gestures, SenseCare emotion AI sensing can accurately sense and classify an individual’s emotional state with a high degree of refinement. SenseCare pushes the strategic envelope when it comes to discerning actual emotions through the use of multimodal techniques.
- Detection techniques based on facial expressions can diagnose and provide effective treatment for patient pain improvement. The effectiveness of a recovery from stroke, acute care, and incisions can be reliably determined by recording facial expressions using a set of facial muscle-based action units. Home Health Aide with detection is highly beneficial when it comes to efficient monitoring of the different stages of recovery.
- Our convolutional approach for feature extraction and classification detects early feature-level and late feature-level. As soon as SenseCare’s facial detection identifies a disorder such as dementia or depression, doctors can recommend an integrated medical care model for the patient.
- Occupational Therapists can use emotion analyses to evaluate improvement in the management of apathy, anxiety, depression, psychosis, dementia, ICD, sleep disturbance, and pain.
- Emotion analyses can be used to create a risk profile in the patient’s database. Doctors can then triage patients according to risk profiles. This enables the optimization of resources in real time so that doctors can focus on those patients most in need.
- The emotion screening tool used in conjunction with instruments for specific symptom assessment serves to strengthen the recognition and holistic management of geriatric care.
- SenseCare’s emotion-sensing algorithm is especially useful in detecting pain in pre-stroke/post-stroke. The system detects pain using facial muscle-based Action Units (AU). It performs the task with deep learning-based Automated Facial Expression Recognition and can jointly detect the complete set of pain-related Action Units. The facial expression is processed to estimate a vector of Action Units using a deep convolutional neural network with concatenation to form a table of AU values. It is especially useful in detecting pain during pre-stroke/post-stroke recovery when patients cannot communicate their pain verbally.